Association for Public Policy Analysis & Management

While the United States is experiencing higher college enrollment rates for all racial and ethnic groups, less progress has been made on college-completion rates with particular challenges in closing the gaps between White students and their underrepresented Hispanic and Black peers. In addition, the nation is experiencing unprecedented demographic changes in the population as well as its school systems with half of all new births and a majority of all students in the public schools categorized as non-white (U.S. Department of Education, 2014; Doughtery & Jordan, 2012; Pew Hispanic, 2014). Using state administrative data from Texas, along with select national datasets, we examine the college-completion gap between underrepresented minority (URM) students and White students, separately, through a variance decomposition analysis.

Through the variance decomposition technique we seek to disentangle the influence of individual characteristics, high school curriculum and context, and the postsecondary institution attended on college completion. In addition, we also account for the types of institutions Hispanic and Black students in Texas are more likely to attend in comparison to their White peers, Hispanic-Serving Institutions (HSIs) and Historically Black College and Universities (HBCUs; Flores & Park, 2013). This technique has previously been utilized to examine factors that contribute to gender wage gaps (Oaxaca & Ransom, 1998; Stanley & Jarrell, 1998) and most recently to study gaps in completion by postsecondary institutional sectors to determine what percentage of the gap is explained by different contextual factors, e.g., two-year versus four-year college attendance (Bound, Lovenheim & Turner, 2010)

We find that pre-college characteristics (a combination of individual and high school context) contribute upwards of 61% of the total variance for both Hispanic and Black students, as compared to their White peers. That is, more than half of the completion gap for URM students is explained by pre-college characteristics. Postsecondary context explains approximately 35% of the total variance, with differing contributions of this portion of the variance attributable to Minority-Serving Institution status, by race. Of greatest relevance, however, is the fact that, while pre-college characteristics explain nearly the same amount of variance in the Hispanic-White and Black-White college-completion gaps, economic disadvantage is the main factor driving the Hispanic-White gap, whereas academic preparation drives the variance explained in the Black-White gap. While these results by no means preclude postsecondary institutions in Texas or in any U.S. state for that matter from addressing the college-completion gap, the data suggest that college completion is not just a postsecondary issue.

These results have strong implications in an era of increased accountability in postsecondary education. This analysis, which is based on a comprehensive K-16 dataset from Texas, indicates that a substantial share of the influence on graduation rates is accounted for by differences in factors that happened to students before college that shape their lives—family resources, academic preparation, community context, etc.—and that the key non-college factors may well be different for different communities of color. Recommendations for other states with similar state administrative data are offered.